On 09/12/2013 10:06 AM, Andrew Hughes wrote:
> Enjoy:
>
> http://cr.openjdk.java.net/~andrew/jdk6-hs23-merge/
>
> Differences that remain between this and jdk7u25-b15 [0]:
Thanks for the patch and explaining the differences. All the changes
looks sensible to me.
> * src/os/linux/vm/os_linux.cpp: keep old version of get_stack_bounds,
> reverting 7017193 which causes
> performance regressions and keeping the simpler fix in [1]
Perhaps we should stick with the upstream version? Is diverging from
upstream really worth it?
> I'll look at merging some of these minor typo changes
> into 8 if not already present.
Please do! It will be great to have the mistakes and typos fixed upstream.
One concern, though (not that it needs to be resolved right now, but I
figured I would point it out anyway):
In current hotspot at jdk6, src/share/vm/runtime/globals.hpp contains
this bit:
experimental(bool, EnableInvokeDynamic, false,
"recognize the invokedynamic instruction")
This disables JSR292 support, if I am reading this correctly. The new
hotspot enables it:
diagnostic(bool, EnableInvokeDynamic, true,
"support JSR 292 (method handles, invokedynamic,"
"anonymous classes")
For compatibility with Java 6, shouldn't this be kept disabled?
Thanks,
Omair
--
PGP Key: 66484681 (http://pgp.mit.edu/)
Fingerprint = F072 555B 0A17 3957 4E95 0056 F286 F14F 6648 4681
